Let $X_1$, $X_2$, $X_3$,$\ldots$ be independent and identically distributed random standard normal random variables.

Define the following: $$Z_n = \sqrt{n} \frac{X_1 + X_2 + \cdots + X_n}{X_1^2 + X_2^2 + \cdots + X_n^2}$$

I am trying to prove that $Z_n$ converges in distribution to a standard normal random variable $Z$.

This looks tantalizingly close to something to do with the Central Limit Theorem, but I cannot quite see how to make the CLT apply here.

I have attempted to show this convergence by use of moment generating functions, but that got very messy very quickly.

Is there a simple way to show the desired result using the CLT, or is it more involved?


Rewrite $Z_n$ as $$Z_n = \frac{X_1 +...+X_n}{\sqrt{n}} \cdot \frac{n}{X_1^2 + ... + X_n^2}= A_n \cdot B_n.$$ Now, $A_n$ follows a standard normal and $B_n \to 1$ in probability by the weak law of large number and the continuous mapping theorem (since $\Bbb E[X_1^2]$).

Therefore, by Slutsky's theorem $Z_n$ converges in distribution to a standard normal.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.